Content uploaded by Marsha K Morgan
Author content
All content in this area was uploaded by Marsha K Morgan on Aug 09, 2016
Content may be subject to copyright.
Design and sampling methodology for a large study of preschool children’s
aggregate exposures to persistent organic pollutants in their everyday
environments
NANCY K. WILSON,
a
JANE C. CHUANG,
b
RONALDO IACHAN,
b
CHRISTOPHER LYU,
a
SYDNEY M. GORDON,
b
MARSHA K. MORGAN
c
HALU
ˆKO
¨ZKAYNAK
c
AND LINDA S. SHELDON
c
a
Battelle Memorial Institute, Durham, North Carolina, USA
b
Battelle Memorial Institute, Columbus, Ohio, USA
c
National Exposure Research Laboratory, USEPA, Research Triangle Park, North Carolina, USA
Young children, because of their immaturity and their rapid development compared to adults, are considered to be more susceptible to the health effects of
environmental pollutants. They are also more likely to be exposed to these pollutants, because of their continual exploration of their environments with all
their senses. Although there has been increased emphasis in recent years on exposure research aimed at this specific susceptible population, there are still
large gaps in the available data, especially in the area of chronic, low-level exposures of children in their home and school environments. A research
program on preschool children’s exposures was established in 1996 at the USEPA National Exposure Research Laboratory. The emphasis of this
program is on children’s aggregate exposures to common contaminants in their everyday environments, from multiple media, through all routes of
exposure. The current research project, ‘‘Children’s Total Exposure to Persistent Pesticides and Other Persistent Organic Pollutants,’’ (CTEPP),isa
pilot-scale study of the exposures of 257 children, ages 11
2–5 years, and their primary adult caregivers to contaminants in their everyday surroundings.
The contaminants of interest include several pesticides, phenols, polychlorinated biphenyls, polycyclic aromatic hydrocarbons, and phthalate esters. Field
recruitment and data collection began in February 2000 in North Carolina and were completed in November 2001 in Ohio. This paper describes the
design strategy, survey sampling, recruiting, and field methods for the CTEPP study.
Journal of Exposure Analysis and Environmental Epidemiology (2004) 14, 260–274. doi:10.1038/sj.jea.7500326
Keywords: preschool children, survey sampling, recruiting, child day care, aggregate exposures, pesticides, persistent organic pollutants
Introduction
Background
Young children, because of their immaturity and their rapid
development, are more likely to be susceptible than are adults
to the health effects of environmental pollutants. As a result
of their continual exploration of their environments with all
their senses Fsight, hearing, touch, smell, and taste Fthey
are also more likely to be exposed to the pollutants. The
possibly greater risk to children from pollutants in their
environments led Congress in 1996 to enact the Food
Quality Protection Act (FQPA, 1996). The FQPA requires
an additional reduction by a factor of 10 in the residues of
pesticides in foods that were previously allowed under the
Federal Food, Drug, and Cosmetic Act (FFDCA) and the
Federal Insecticide, Fungicide and Rodenticide Act (FIF-
RA). The FQPA also requires that aggregate exposures,
from all media with which children might come in contact,
through all routes of exposure (inhalation, ingestion, and
dermal absorption) be considered in setting the allowable
residues.
The increased emphasis on children’s exposures to
pesticides and other organic pollutants has led to a surge in
recent years in the number of research studies aimed at this
specific susceptible population (Duggan et al., 1985; Lewis
et al., 1994; Stanek and Calabrese, 1995; Heil et al., 1996;
Wilson et al., 1996, 1999, 2000, 2001, 2003; Chuang et al.,
1998, 1999a; Gurunathan et al., 1998; Mukerjee, 1998;
Vonmanikowsky et al., 1998; Landrigan et al., 1999; Adgate
et al., 2000, 2001; Akland et al., 2000; Fenske et al., 2000a, b,
2002a,b; Landrigan, 2001; Nishioka et al., 2001; O’Rourke
et al., 2000; Freeman et al., 2001; Freedman et al., 2001;
Heudorf and Angerer, 2001; Karmaus et al., 2001; Lu et al.,
2001; Mills and Zahm, 2001; Rigas et al., 2001; Baker et al.,
2002; Brock et al., 2002; Buckley et al., 2000; Koch et al.,
2002; Perera et al., 2002; Raymer et al., 2002; Tulve et al.,
2002a, b; Wilhelm et al., 2002). However, large gaps in the
available data still exist (Cohen Hubal et al., 2000a, b;
Needham and Sexton, 2000), especially in the area of
Received 22 May 2003; accepted 4 September 2003
1. Address all correspondence to: Dr. Nancy K. Wilson, Battelle, 100
Capitola Drive, Suite 301, Durham, NC 27713-4411, USA. Tel.: þ1-919-
544-3717 ext. 140. Fax: þ1-919-544-0830. E-mail: wilsonk@battelle.org.
Journal of Exposure A nalysis and Environmental E pidemiology (2004) 14, 260–274
r2004 Nature Publishing Group All rights reserved 1053-4245/04/$25.00
www.nature.com/jea
chronic, low-level exposures of children in their home and
school environments. Additionally, the interesting question
of whether children and adults who spend most of their time
in the same household microenvironment have similar
exposures and potential doses has not been investigated
extensively.
To address some of these issues, EPA’s National Exposure
Research Laboratory in 1998 began designing the large-scale
pilot study, ‘‘Children’s Total Exposures to Persistent
Pesticides and Other Persistent Organic Pollutants
(CTEPP),’’ which was approved for implementation by the
US Office of Management and Budget in February 2000.
The CTEPP study is a study of the aggregate exposures of
257 preschool children and their primary adult caregivers to a
suite of about 50 persistent organic pollutants in their
everyday environments. The field portion of the CTEPP
study began in February 2000 and was completed in
November 2001 (Lyu et al., 2002). This paper describes
the design strategy, survey sampling, recruiting, and field
methods for CTEPP.
Design Considerations
Since the chronic exposures of preschool children have been
studied less extensively than those of older persons, the
emphasis of the CTEPP study is on children in the age range
11
2–5 years. The three major objectives of the study are: (1) to
measure the chronic, aggregate (total) exposures of approxi-
mately 260 preschool children and their adult caregivers to
low levels of a suite of persistent pesticides and other
persistent organic pollutants that the children may encounter
in their everyday environments, (2) to apportion the routes of
exposure, and estimate the relative contributions of the
exposure media, and (3) to identify important hypotheses to
be tested in future research.
Within the framework of the above objectives, there are
several hypotheses that can be tested within this group of
children using the CTEPP data. The first set of hypotheses is
whether the children’s exposures are the same (a) at home
and at day care, (b) in low-income and middle/upper-income
households, (c) in urban and rural environments, (d) as those
of adults in the same households, and (e) through different
routes for different chemical classes.
The second set of hypotheses is whether, considering the
dominant routes of exposure, (a) ingestion is a major route of
exposure, and (b) diet is the major contributor to the
children’s ingestion exposures.
Preliminary studies of microenvironmental concentrations
of a large number of persistent organic pollutants in several
child day care centers (Wilson et al., 2001) and the aggregate
exposures to these same pollutants of nine children who
attend child day care (Wilson et al., 2003) provided tests of
the field sampling and analysis methods, initial estimates of
the ranges of data that might be expected, and information
necessary to develop a sampling strategy.
To acquire data in CTEPP for a broad range of children’s
possible microenvironmental exposures, the children were
from urban, rural, low-income and middle/upper-income
households in several distinct geographic regions. Half of
these children attended child day care centers and half did not
attend day-care, but stayed at home with an adult caregiver.
One child per household participated in the study. A parent
or other adult, who was the primary caregiver in the
household for the selected child, also participated. To reduce
costs, the children were chosen from two states, North
Carolina (NC) and Ohio (OH), in which there are major
EPA research facilities.
A broad range of compounds, listed in Table 1, was chosen
for measurement in environmental media to provide
aggregate exposure data that might also be useful for
estimating cumulative exposures to particular compound
classes. These compound classes are: polycyclic aromatic
hydrocarbons (PAH), polychlorinated biphenyls (PCB),
organochlorine insecticides (OC), organophosphorus insecti-
cides (OP), pyrethroid insecticides (PY), phthalate esters
(PE), and phenols (PH), the acid herbicides 2,4-D, dicamba,
and 2,4,5-T, and the triazine herbicide atrazine. Additionally,
several hydroxylated metabolites of the PAH, the chlorpyr-
ifos metabolite 3,5,6-trichloro-2-pyridinol (TCP), the diazi-
non metabolite 2-isopropyl-4-methyl-6-hydroxypyrimidine
(IMP), the pyrethroid metabolite 3-phenoxybenzoic acid
(3-PBA), and unchanged pentachlorophenol and 2,4-D were
chosen for measurement in urine.
The complete study required development of Standard
Operating Procedures (SOPs), covering subject recruitment,
field sampling, and storing and shipping of samples and data
collection forms. It also required development of SOPs
covering data processing, laboratory procedures, and the
study database, to total 50 SOPs altogether. This paper
describes the design and field sampling aspects of the CTEPP
study.
Methods
Statistical Power
Based on our findings in earlier studies of the PAH exposures
of 24 children in low-income families (Chuang et al. , 1999b)
and in our preliminary small studies of children’s exposures
to persistent organic pollutants (POP) in child day-care
centers and at home (Wilson et al., 1999, 2000, 2001, 2003),
the numbers of children necessary in the CTEPP study to
achieve defined statistical powers were calculated. The data
from the above studies indicated that the POP concentrations
tend to be log-normally distributed in environmental media,
with standard deviations of the log (POP concentrations)
ranging generally from 0.50 to 2.0. The differences in
geometric mean POP concentrations between low-income
and higher-income families as well as between day-care
Design and sampling methodogy for the CTEPP study Wilson et al.
Journal of Exposure Analysis and Environmental Epidemiology (2004) 14(3) 261
centers and homes ranged between 0% and 500%; between
urban and rural areas they ranged from 0% to 250%. For
concentrations of seven target POP in six sampled media, the
median percent difference in POP concentrations between the
groups in the above small studies was 60%, and the median
standard deviation in log-transformed POP concentrations
was 1.0.
Power calculations were performed with the following
assumptions: (1) a two-sample t-testatthe5%significance
level on log-transformed POP concentrations to compare
POP exposures of groups of children in low-income vs.
middle/upper-income families, in child day-care centers vs.
homes, and in urban vs. rural areas, (2) sample sizes to
provide 80–90% power for detecting a significant difference
between the groups when the actual percent difference ranges
from 10% to 200%, and (3) sample sizes assuming standard
deviations of the log-transformed POP concentrations of 0.5,
1.0, 1.5, and 2.0.
The results of these calculations indicated that if the
standard deviation of log-transformed POP concentrations is
1.0, and the actual percent difference in POP exposures
between any two groups of children is 50%, then a sample
size of approximately 100 children per group would provide
80% power for detecting a statistical difference in POP
exposures between the groups. A sample size of approxi-
mately 130 children per group would provide 90% power for
detecting the difference.
To test the CTEPP hypotheses with at least 80% power,
and to allow for missing samples and other data losses, at
least 120 children each from low-income and middle/upper-
income families would participate. Ideally, these children
would be distributed evenly in each group: day-care vs.
nonday-care, urban vs. rural, and low vs. middle/upper-
income, to achieve the best statistical power for comparisons.
US Census data for the states of North Carolina (NC) and
Ohio (OH) indicate that their 1999 poverty rates were 10.5–
12.5%, near the US average of 11.5%. The state populations
were 60.7% urban and 29.8% rural in NC, and 77.3%
urban and 22.7% rural in OH (US Census Bureau, 2002).
Therefore, to obtain a probability-based stratified sample, the
group of low-income subjects would be smaller than the
group of middle/upper-income subjects, and the rural group
would be smaller than the urban group. If, in order to test the
urban/rural and income level hypotheses, these smaller
groups were extremely oversampled to provide approxi-
mately equal numbers in each group, the effect would be
imbalance in sampling rates and weights, and inefficient
sampling design. Therefore, to increase the representativeness
of the CTEPP study, allow intergroup comparisons, and
meet the study objectives, yet stay within the constraints of
Ta b l e 1 . Target compounds for the CTEPP Study
Target compounds Number of
compounds
Reason for selection
Four-ring and larger polycyclic aromatic hydrocarbons ( PAH) 9 Some probable human carcinogens (B2 carcinogens), possible
endocrine disrupters; ubiquitous combustion products
Phthalates: benzylbutyl and di-n-butyl phthalate 2 Possible endocrine disrupters, possible carcinogens; widely used
plasticizers
Phenols: pentachlorophenol, nonylphenols, and bisphenol-A 3 Possible endocrine disrupters, teratogens, carcinogens; widely
used wood preservative, pesticides, surfactants, and plasticizers;
bisphenol-A in polycarbonate plastics, baby bottles, dental
sealants, and coatings.
Organochlorine pesticides: lindane, aldrin, alpha- and gama-chlordane,
p,p0-DDT, p,p0-DDE, dieldrin, endrin, heptachlor, and
pentachloronitrobenzene
10 Possible endocrine disrupters, toxicity, neurotoxicity,
developmental neurotoxicity, some probable carcinogens;
former widespread use, both indoors and outdoors; some still
approved for specific uses, e.g., lindane in head lice shampoos
Polychlorinated biphenyls (PCB): penta-, hexa-, and hepta-chlorinated
biphenyls
17 Possible endocrine disrupters, developmental neurotoxicity;
coplanar PCB are possible carcinogens; former widespread
industrial use
Organophosphorus pesticides: diazinon and chlorpyrifos 2 Possible endocrine disrupters; acetyl cholinesterase inhibitors;
household insecticides, may persist indoors; dietary residues
common
Acid herbicides: 2,4-dichlorophenoxyacetic acid (2,4–D), dicamba, and
2,4,5-trichlorophenoxyacetic acid (2,4,5-T)
3 Possible endocrine disrupters; home lawn herbicides
Triazine pesticide: atrazine 1 Carcinogenicity, possible endocrine disrupter; found extensively
in drinking water in the mid-west US
Pyrethroid pesticides: cis/trans permethrins and cyfluthrin 2 Neurotoxicity, possible endocrine disrupter
Urinary metabolites: hydroxylated PAH, 3,5,6-trichloro-2-pyridinol,
2-isopropyl-4-methyl-6-hydroxypyrimidine, and 3-phenoxybenzoic acid
4 Improve exposure and dose estimates
Unmetabolized target compounds in urine: pentachlorophenol and 2,4-D 2 Improve exposure and dose estimates
Design and sampling methodogy for the CTEPP studyWilson et al.
262 Journal of Exposure Analysis and Environmental Epidemiology (2004) 14(3)
reasonable expenditures of resources, the following sampling
plan was developed. This plan takes into account the
necessary compromises between the aforementioned conflict-
ing statistical goals.
Survey Sampling
The probability-based, multistaged stratified random sam-
pling plan outlined in Figure 1 was devised to recruit
participants to the CTEPP study. This plan enabled
acquisition of the data needed to meet the objectives of the
CTEPP study regarding aggregate exposure measurement,
apportionment of routes of exposure, and hypothesis testing.
It increased the representativeness of the participants by
using two sample frames and maximized the amount of
useful information that can be obtained, while balancing the
conflicting demands of representativeness and hypothesis
testing.
The target population of the CTEPP study was children
between the ages of 11
2–5 years who resided in NC and OH.
To allow hypothesis testing at 80% power, and to allow for
missing samples and other possible data losses, approxi-
mately 260 children (and their caregivers) were targeted to
participate, with 130 originating from each state. The sample
was stratified so that half of the children targeted in each state
attended child day-care centers, and half did not.
In the first sampling stage, six counties in each of the two
states were selected by stratified random sampling, with the
strata determined by region within the state and the degree of
urban character (urbanicity). The regions in each state
corresponded to three distinct geographical areas, as
illustrated in Figure 2. In NC, the geographical areas were
the coastal plain, the Piedmont, and the mountains. In OH,
they were the northern, middle, and southern regions. The
urbanicity stratification was imposed by classifying counties
within each region as predominantly urban or predominantly
rural, based on US Census definitions. Within each of the
region and urbanicity strata, counties were selected with
probability proportional to the size (PPS) of their low-
income population, which was also essentially proportional
to the total population of the counties. Here, the low-income
population was defined following the federal guidelines for
assistance under the Women, Infants, and Children (WIC,
2000) program; that is, household income less than 185% of
the federal poverty guidelines (Federal Register, 2000). In
2000, the WIC eligibility guideline for a family of two was
$20,813 and for a family of four was $31,534.
Following the first sampling stage (of counties), the
sampling plan branched into a day-care component and a
telephone survey component, as shown in Figure 1. Within
the day-care component, the next sampling stage involved
selecting day-care centers. For each state, a sampling frame
of all state-licensed day-care centers in the six selected
counties was constructed. Home-based centers, part-time
centers, and centers serving only school-aged children were
excluded. The sampling frame was then stratified by county
and by whether or not the center received Federal assistance
to serve low-income clients (Head Start centers). Within each
stratum, a targeted number of day-care centers were selected
with probability proportional to the total number of children
enrolled in the center. Approximately 16 day-care centers
were targeted for selection in each state, with at least four of
these being Head Start centers.
The final sampling stage of the day-care component
involved selecting a random sample of eligible children from
up to two classrooms in the selected centers. A child
attending the center was deemed eligible for the study if the
child was within the age range of the study’s target
population, was toilet-trained or able to provide at least
one urine sample, was not being breast-fed, and attended the
center at least three consecutive days per week, for at least
25 h/week. Approximately six children were targeted for
selectionineachoftheHeadStartcenters,whilefour
children were targeted in each of the other centers. This
allowed for oversampling of low-income children who
attended day care. Each selected child was classified as
belonging to a low-income or a middle/high-income family,
using the low-income criteria given above.
The telephone survey component of the sampling plan
involved selecting households with children who did not
attend a day-care center. A stratified random sampling plan
was adopted for the telephone survey, with the strata
corresponding to the six counties. Initial contacts with
Two State Sampling Plan (NC and OH)
Select Representative Counties – Six per State
(4 urban, 2 rural)
NC Regions: Mountain, Coastal, Piedmont
OH Regions: South, Central, North
Two Sampling Frames in Each State
Sample Component
Telephone Screening
Sample Component
Day Care Centers
Use List-Assisted Telephone
Sampling Technique;
Recruit Age-Eligible Children
Stratify by Income Level;
~ 64 Participants per State
1st Stage
Select ~ 16 Day Care Centers with
Probability Sampling
2nd Stage
Randomly Select Age-Eligible Children
~ 64 Participants per State
Figure 1. Overview of the sampling plan for the CTEPP study.
Design and sampling methodogy for the CTEPP study Wilson et al.
Journal of Exposure Analysis and Environmental Epidemiology (2004) 14(3) 263
selected households involved a screening process, discussed in
greater detail in the recruiting section, which determined
whether an eligible child resided in the household and how
the household was categorized based upon income level. The
eligibility criteria for the telephone sample were the same as
for the day-care sample, except that the child did not attend
day care. In this way, targets placed on numbers of
households to recruit within each income level could be
monitored during the recruitment process. As in the day-care
center component, the sampling plan for the telephone survey
provided for oversampling of low-income households.
Theanticipatedsamplesizewas128childrenineachstate,
with half originating from the day-care center sample
(children who attend day care) and half from the telephone
survey (children who do not attend day care). This dual
frame approach provided maximum coverage for the target
population.
In summary, the survey sampling plan allowed general-
ization of the sample data to a wider and more diverse area in
each of the two states so that valid statistical inferences could
be made. The design also simplified the sampling frame
construction and sample selection process. This sampling
Figure 2. (a) The six NC and (b) the six OH counties selected for the CTEPP study. (Maps adapted from state maps available at http://
quickfacts.census.gov/.)
Design and sampling methodogy for the CTEPP studyWilson et al.
264 Journal of Exposure Analysis and Environmental Epidemiology (2004) 14(3)
plan and the supporting power calculations are described in
detail elsewhere (Chuang et al., 1998).
Recruitment
Recruitment of NC participants began in February 2000,
although a 4-month hiatus was required by the US Census
from March through June, and continued from July through
December 2000 in the four mountain and Piedmont counties,
and in the two eastern counties from February through
March 2001. In Ohio, recruitment took place in January
through November 2001.
Although the principal subjects in the CTEPP study were
preschool children, a significant burden was placed on the
parents or primary caregivers, additional household mem-
bers, and teachers for collection of information and physical
and biological samples. Parents and teachers provided
information on classroom and household characteristics,
including pesticide use and food consumption and prepara-
tion; collected food, hand-wipe, and urine samples; and kept
child time-activity diaries over 2 days. Study staff collected
additional physical and observational information, such as
general neighborhood and housing conditions, nearby
industries or other potential sources of pollution, and
household/classroom floor plans.
Recruitment strategies included minimizing the burden on
participants, ensuring confidentiality, providing incentives for
participation, and using carefully selected and well-trained
field staff. Throughout, the staff were encouraged to be
sensitive to participants’ concerns and to persevere in
recruitment.
The most frequent concern related to participant burden
waslackofstaffbycentersorlackoftimebyparents.Day-
care teachers in particular were concerned about collection
and storage of urine samples. Several ways of reducing
participant burden included providing individual training to
participants prior to the field sampling, providing assistance
for urine collection at the centers, offering flexible sampling
schedules, and providing a project toll-free telephone number
to call for assistance. Additionally, actual contact time
between staff and participants during sampling was kept as
short as possible.
A major worry for some participants, especially the
directors and staff of child day-care centers, was whether
individual data would be released to any regulatory agency or
to others. To allay this concern, a Certificate of Confidenti-
ality for the study was obtained from the National Institute
of Mental Health. This Certificate provides legal protection
of the privacy of the individual data. Under this Certificate,
the study researchers cannot and will not release any
individual data to anyone, including the courts, without
written permission of the individual.
To encourage participation, both monetary and nonmo-
netary incentives were offered to participants. Participating
families and child day-care centers received $100 to cover
their costs of providing food and other samples. If the
children were to be videotaped for about 2 h, an additional
incentive payment of $50 was furnished to the participating
household; a $25 gift certificate for a book or other
appropriate item for the classroom was provided to childcare
centers. At each visit to homes or centers, field staff brought
small age-appropriate gifts for the participating children.
Field staff encouraged participants to realize that they were
performing important research, and that their participation
was valuable. Participants were given a project T-shirt and
pen. All participants received a framed certificate, acknowl-
edging their contributions, at the conclusion of field
sampling. At the end of the study, participants also received
a brief study summary report.
To enhance response rates in the CTEPP study, user-
friendly materials and brochures were developed. Letters and
statements of endorsement were obtained from childcare
organizations, such as the National Head Start Organiza-
tion, and from past pilot study participants. Press releases
about the study from the US EPA were used in the selected
areas, and the EPA principal investigator provided radio
interviews. Prior to personal contact with centers and
parents, introductory letters and brochures were sent to
them by overnight courier (FedEx). Study staff made
multiple follow-up calls and personal visits to potential
participants. Throughout, the study staff tried to develop a
sense of a research partnership between centers, teachers,
parents, and researchers.
For the initial telephone screening of potential participants,
scripts were developed for interviewers, so that the screening
information could be entered directly into the computer.
Written consent forms for participation and for possible
future contact were developed. Additionally, to allow
systematic collection of information necessary for calculation
of aggregate exposures and interpretation of the measure-
ments, the following forms were developed: recruitment
survey, household and day-care center observation surveys;
premonitoring questionnaires for parents and centers; post-
monitoring questionnaires for day-care centers, day-care
parents, and nonday-care parents; and child time–activity
diaries and food surveys for day-care teachers, day-care
parents, and nonday-care parents.
Recruitment of Day-care Participants Recruitment of
participants in the day-care center component of the study
was conducted in two stages, as shown in Figure 3. In the
first stage, from a complete list of child day-care centers in the
selected NC and OH counties, 16 were selected randomly in
each state. If some refused, additional centers were selected
randomly. In NC, of 32 centers, 22 were successfully
contacted; five were ineligible, and four refused. In OH, of
33 centers, 28 were successfully contacted, four were
ineligible, and eight refused. Introductory letters and a
study brochure were sent by overnight courier to the center
Design and sampling methodogy for the CTEPP study Wilson et al.
Journal of Exposure Analysis and Environmental Epidemiology (2004) 14(3) 265
directors, along with a $25 toy store gift certificate for the
center, and a request for an interview. During the personal
visit with the director by a CTEPP staff member, the study
was explained further and informed consent for the center’s
participation was solicited. In the second stage, eligible
children were identified, and four to six children were
randomly selected in each center. Introductory letters and a
study brochure were sent to the parents of the selected
children, followed by meetings at the center and personal
visits to explain the study further and obtain informed
consent.
Recruitment of Nonday–care Participants Recruitment of
participants in the nonday-care component (telephone
sample) of the study was conducted using a two-phase
sample that combined a list sample of households with listed
telephone numbers with a list-assisted random digit dialing
(RDD) sample, as shown in Figure 4. To identify efficiently
those households having one or more children in the eligible
age range and meet the sampling targets in the income
domains, potential nonday-care participants were identified
using commercially available directories (Marketing Systems
Group, Genesys Sampling System, http://www.genesys-
sampling.com) of households in the selected geographical
areas and household income strata. All directory-listed
households were assigned to one of four strata: (1) income
above $25,000 with one or more children in the target age
range, (2) income below or equal to $25,000 with one or
more children in the target age range, (3) income above
$25,000 with no children in the target age range, and (4)
income below or equal to $25,000 with no children in the
target age range.
To ensure inclusion of households that did not appear in
the directories, an RDD approach was used. All telephone
exchanges in the selected county were identified. Those
having very low percentages of directory-listed households
(primarily nonresidential/business areas) were deleted. From
the remaining exchanges, a systematic random sample of all
numbers was drawn. The samples, corresponding to the four
strata above, and the RDD samples were combined in
replicate files provided by the sampling firm. Households
without home telephones were partially represented in the
day-care group; not all such children attend day care.
Purchase telephone sample list for the target counties
Send introductory letters to households on the list
Conduct telephone screening calls
Screen for eligibility, using CATI
Conduct follow-up calls and schedule visits
Conduct follow-up calls for refusing parents in an attempt to
reverse refusal
Conduct follow-up visits
Meet with parents, explain the study, and
obtain informed consent or obtain mail-back consent
Conduct random selection for final participants
Figure 4. Nonday-care (telephone sample) recruitment.
First Stage Second Stage
Obtain complete list of
day care centers in the
target counties
Randomly select 16 day
care centers from the list
in each state
Call the selected day care
centers to confirm address
and name of director
Collect day care
information: number and
age of eligible children
per class
Mail introductory letters to
the directors of selected
day care centers
Conduct follow-up calls
and schedule visits
Conduct follow-up visits
and obtain informed
consent or
obtain mail-back
consent
Identify and list
eligible children
Randomly select 4-6
children in each day
care center
Send introductory letters
to the parents of
selected children
Conduct follow-up calls
and schedule visits
Conduct follow-up visits.
Meet with the parents at
the day care center,
explain the study, and
obtain informed consent
or obtain mail-back
consent
Conduct follow-up
calls or visits for
refusing centers by
study survey director
in an attempt to
reverse the refusal
Conduct follow-up
calls or visits for
refusing parents by
study survey director
in an attempt to
reverse the refusal
Figure 3. Child day-care center recruitment.
Design and sampling methodogy for the CTEPP studyWilson et al.
266 Journal of Exposure Analysis and Environmental Epidemiology (2004) 14(3)
Introductory letters and a study brochure were sent to
households in the files that had valid addresses. All numbers
in the files were called and screened for eligible subjects. The
final participants were selected randomly from the eligible
subjects. Staff visited personally those households that agreed
tentatively to participate. At these visits the staff explained
the study further and obtained informed consent.
Field Sampling Methods
To obtain data that would allow estimation of the children’s
aggregate exposures, and those of their primary caregivers, to
the large variety of organic pollutants targeted in this study,
several types of environmental media samples were collected
in the households and child day-care centers Findoor and
outdoor air, play area soil, and carpet dust or if no carpet,
hard floor surface wipes. If pesticides had been applied in the
7 days prior to sampling, transferable residues, hard floor
surface wipes in addition to carpet dust, and food prepara-
tion surface wipes were also collected. Additionally, personal
samples of liquid and solid food, drinking water, hand-wipes,
and urine were collected for the individual participants. Other
field sampling activities included pre- and postmonitoring
interviews, house/building characteristic observation survey,
collection of a child/adult activity and food diary, collection
of day-care food menus, and videotaping the activities of 26
children.
For the day-care component, field sampling took place
simultaneously during a 48-h period at each child’s day-care
center and at her/his home. All selected children in a given
center classroom were sampled during the same week. For
the telephone component, sampling took place over a 48-h
period, scheduled at the convenience of the parents, but with
consideration given to efficient use of staff and minimizing
travel time.
Environmental Samples Indoor and outdoor air, carpet
dust, and play area soil samples were collected by the
methods delineated in our earlier studies (Chuang et al.,.
1987, 1991, 1994, 1995, 1998, 1999b; Wilson et al., 1991,
2001, 2003). Indoor and outdoor air at day care and at home
were sampled for 48 hr at 4 l/min with a URG-2000 sampling
cartridge (University Research Glassware Corp., Chapel
Hill, NC, USA) followed by an SKC Model 224-PCXR8
pump indoors (SKC, Inc., EightyFour PA) or a Thomas
Model 107CAB18A pump outdoors (Thomas Compressor
and Vacuum Pumps, Sheboygan, MI, USA). The 10 mm
impactor-equipped inlet was followed by a glass cartridge
containing a quartz fiber filter followed in series by XAD-2
resin, to collect targeted compounds in both the particulate
and vapor phases (Chuang et al., 1987, 1991, 1999b; Wilson
et al., 1991, 1996). The indoor air sampling SKC pump was
placed in a Styrofoam box equipped with a cooling fan, and
then in a playpen with a net over the top to keep curious
children from making adjustments to the equipment. The
outdoor air sampling Thomas pump and controls were
sheltered in a plastic dog house, purchased for the study,
which was located centrally in the children’s outdoor play
areas. The air sampler inlets were placed at the approximate
breathing height of the children, about 75 cm from the floor
or ground.
Classroom and house carpet dust samples were collected
with a High Volume Small Surface Sampler (HVS3; Cascade
Stack Sampling Systems Inc., Bend, Oregon) using a
standard ASTM method (Chuang et al., 1994; Lewis et al.,
1994; ASTM, 1997) in the areas indicated by the teacher or
parent as being where the children played most often. If there
was no carpet, or sufficient carpet sample could not be
obtained, a hard floor surface wipe sample (see below) was
collected. Play area soil was scraped from the top 0.5 cm of
exposed soil over a 0.1 m
2
area in the location identified by
the teacher or the parent as most often used by the children
(Chuang et al., 1995). The selected outdoor play areas were
usually bare of grass.
At homes and day-care centers that had indoor or outdoor
pesticide applications within the 7 days preceding the
sampling period, additional environmental samples were
collected. Transferable residues were collected from
24 cm 1 m areas of carpeted and uncarpeted floor surfaces
in three locations pointed out by the parent or teacher, as
those where the child played most, by the polyurethane
(PUF) roller method (Lewis et al., 1994; Camann et al.,
1996; Fortune, 1997). Hard floor surface residue samples
were collected by wiping twice in opposite directions, with a
precleaned 10 10 cm
2
3-ply gauze pad (Sof-Wick, Johnson
and Johnson) moistened with 2 ml of 75% isopropanol/
water, a 38 38 cm
2
area where the child spent the most
time. Food preparation surface wipe samples were collected
similarly from a 38 38 cm
2
area, on the counter or other
surface where the adult participant usually prepared food.
At least one field blank sample was collected in each
sampled medium at each location.
Personal Samples The children’s parents/caregivers/
teachers were trained by our staff to collect the diet, hand-
wipe, and urine samples. Duplicates of all foods and
beverages eaten during the 48-h sampling period by the
children, and by the nonday-care adult caregivers, were
collected. Duplicates (Acheson et al., 1980; Block, 1982;
Fennema and Anderson, 1991) of the daily meals served to
the children in the participating classrooms and at home were
collected on each of the two sampling days. The same menus
were served to all the children in a given classroom. Menus at
home varied considerably, as expected. If a child brought his/
her food from home, rather than eating food served at day
care, a duplicate of this food from home was collected. Foods
were collected each day as composites of food eaten at home,
and composites of food eaten at day care. The teachers and
caregivers were asked to remove the inedible parts of the
Design and sampling methodogy for the CTEPP study Wilson et al.
Journal of Exposure Analysis and Environmental Epidemiology (2004) 14(3) 267
foods before putting them into the sample containers. Liquid
and solid foods were collected separately in glass jars, 1 gal
for the adult and 1
2gal for the child, and refrigerated until
they were shipped on dry ice to the laboratory for analysis.
Drinking water samples were collected at each home and
day-care center by the field staff.
Each child’s hand-wipe samples, from the entire surface
area of all fingers and the front and back of both hands, and
those of adults in the nonday-care group, were collected by
the participating caregivers according to a standard proce-
dure. The 10 10 cm
2
, 3-ply wipes were of precleaned cotton
gauze (Sof-Wick, Johnson and Johnson), moistened with
2 ml of 75% isopropanol/water. Two wipes for each child
were collected at the day-care center or at home, one just
before lunch and before washing the child’s hands, on each of
the two sampling days. Two additional wipes were collected
at home, just before dinner and before washing the child’s
hands, on each of the two sampling days. Adult hand-wipes
for the day-care group were collected once each day, before
dinner. For the nonday-care adults, hand-wipes were
collected twice each day, before lunch and before dinner.
Urine samples from the participants were collected at times
spread throughout the day, to allow estimation of the average
daily urinary excretion of the target compounds. Because the
toxicokinetics in children of these compounds are not well
known, no attempt was made to time the urine collections
relative to the environmental samples. Each child’s caregiver
at home, or teacher while the child was at day care, collected
the child’s urine samples, which were obtained by having the
child urinate into a plastic bonnet, labeled with the individual
child’s name, inserted under the toilet seat. For each
urination, the contents of the bonnet were poured into a
separate 120 ml plastic bottle, labeled with the individual
child’s name, and refrigerated. The nonday-care adult
caregivers also provided their own urine samples. Each child
in the study supplied three urine samples on each sampling
day (six samples total). Adult caregivers in the nonday-care
group also supplied three samples each sampling day; those
in the day-care group supplied only two. For the day-care
group, four urine samples were collected at home for
each child and caregiver: one in the morning (the first
morning void) and one in the evening before bedtime on
each sampling day. Two additional urine samples were
collected for each child at day care, after lunch on each
sampling day. For the nonday-care group, six urine samples
were collected at home for each child and caregiver: one in
the morning, one after lunch, and one in the evening on each
sampling day.
The wipe and urine samples were stored in chilled coolers
at the centers or at home until collected by our staff. Food
and beverage samples were stored in refrigerators at the
sampling sites until collected by the staff. All samples,
including air and dust samples, were then stored in freezers at
101C or below, until they were shipped on dry ice to the
laboratory for analysis. At the laboratory, all samples were
stored at 801C.
All sample containers, questionnaires, diaries and surveys
were labeled with bar codes with the participant and sample
ID before sampling, and the corresponding field data logs
were also bar-coded, so that a continuous chain of custody
and positive identification was kept for all samples, from
shipment of empty containers to the field, through sampling,
shipment to the laboratory, and thereafter through analysis.
Bar-code labeling assisted in tracking samples and reduced
the amount of staff time required in filling out forms.
Questionnaires and Diaries Separate but similar
questionnaires were used for child day-care centers and for
households. In the premonitoring interview, information was
collected on center and household characteristics, including
children’s ages; parents’ occupations, education, income, and
smoking; carpet and cleaning habits; heating and air
conditioning; drinking water sources; pets; and children’s
usual activities. The field staff obtained height, weight, and
hand tracings for calculation of hand surface area, for both
the child and the primary caregiver.
The field staff did the survey of the house/center building.
It included observations on the floor plan, materials, and
condition of the building, and the nature of the surrounding
area (industrial, residential, etc). A sketch of the floor plan of
the house/building diagrammed the child’s preferred play
area and the locations of samplers and sampled sites.
In the postmonitoring questionnaire, information was
collected on the use of, or child’s contact with, various
potential pollutant sources, such as household chemicals or
pesticides, smoking, and grilling, during the 48-h sampling
period. Detailed information was collected on food prepara-
tion methods, pesticide use, and other activities.
The detailed child activity diary and food survey, which
was filled out by the teacher or parent/primary caregiver,
covered the entire 48-h period. It included what, when, and
where the child ate, slept, and played. Additionally, the
child’s activity levels (active play, quiet play, sleeping), the
nature of the surfaces on which the activities took place, and
the type of clothing worn were recorded over the sampling
period. The food survey collected detailed information about
the child’s current and past eating habits over the preceding
year. Day-care centers were asked to provide copies of the
menus listing the food served to the children for the past
month (or longer, if the menus were available). For the
nonday-care component, the above food survey information,
except menus, was collected for the child participant.
Videotaping Children’s Activities To supplement and to
aid evaluation of the children’s time-activity diaries,
approximately 20% of the children among the OH
participants (10% total CTEPP) were videotaped for about
2 h at their homes. At the premonitoring visit, which was held
Design and sampling methodogy for the CTEPP studyWilson et al.
268 Journal of Exposure Analysis and Environmental Epidemiology (2004) 14(3)
a few days before the actual field sampling, parents/caregivers
were asked for permission to videotape the children during
the first day of field sampling. Consent was easily obtained
from many participants, with consent given by about 40% of
those asked. Eventually, 26 children were videotaped in
various activities, including active and quiet play, indoors
and outdoors.
Results and Discussion
Response Rates
In NC, CTEPP enrolled 13 child day-care centers, completed
sampling activities with 63 households, and thus achieved
98% of the planned target of 64. In OH, the project enrolled
16 child day-care centers, completed sampling activities with
58 households, and thus achieved 91% of the planned target
of 64 participating households. Of 22 NC centers successfully
contacted, five were ineligible, four refused, and 13 agreed to
participate. In OH, four centers were ineligible, eight refused,
and 16 agreed to participate. Including all centers to which
contacts were initially attempted, and treating eligibility
status unknown as refusal, the day-care center response rate
was 53% in NC and 57% in OH. If one were to assume that
the distribution of eligibility and response for centers whose
status is unknown is the same as that for centers whose status
is known, the estimated response rates would be 76% in NC
and 67% in OH.
For day-care families in NC, of 111 successfully contacted
through the participating centers, 26 were ineligible, 16
refused, and 69 agreed to participate; in OH, eight families
were ineligible, 29 refused, and 71 agreed to participate.
Including all day-care parents to whom contacts were
attempted, and treating eligibility status unknown as refusal,
the response rates were 50% in NC and 31% in OH. The
eligibility rates for the day-care sample households were 77%
(85/111) in NC and 93% (100/108) in OH. If one were to
assume that the distribution of eligibility and response for
day-care households whose status is unknown is the same as
for households whose status is known, the estimated response
rates would be 81% in NC and 71% in OH. In NC, a total
of 12,262 telephone numbers were called, and 272 potentially
eligible households were screened and agreed to be visited,
with an eligibility rate of 5%. Of these 272 households, 35
refused and 36 were determined ineligible after visit. A total
of 67 randomly selected nonday-care households participated
in NC, which is 105% of the recruitment target of 64. In OH,
a total of 10,179 phone numbers were called, and 165
potentially eligible households were screened and agreed to be
visited, with an eligibility rate of 4%. Of these 165 eligible
households, 14 refused and 16 were determined ineligible
after visit. A total of 69 randomly selected nonday-care OH
households participated in the study, which is 108% of the
recruitment target of 64.
The response rates for the telephone sample were 58% in
NC and 57% in OH, treating eligibility status unknown as
refusal to participate. If one were to assume that the
distribution of eligibility and response for potential partici-
pants whose status is unknown is the same as that for subjects
whose status is known, these estimated response rates would
be 71% in NC and 83% in OH.
The most frequent reason for refusal to participate, given
by both day-care centers and parents, was lack of time on the
part of teachers or parents. Since the CTEPP study was
designed to elicit abundant detailed information on aggregate
exposures and related ancillary information, which could be
used in broad areas of future research on human exposure, it
put a high burden on the individual teachers and caregivers,
estimated at more than 6 h per adult participant in time alone
during the sampling period. In future studies, care should
continue to be taken to minimize this burden and to ensure
that the necessary compromise between burden and data
requirements is not weighted too heavily on the side of
research needs. Development of appropriate low-burden
exposure measurement methodologies should be encouraged.
Some potential participants expressed concern that the
data might somehow be made public, despite assurances of
complete confidentiality. Some just were not interested. A
problem that undoubtedly affected the center response rates
in NC adversely was the slow recovery of the two selected
counties in eastern NC from flooding caused by Hurricane
Floyd, the burden from which was given by potential
participants as a major reason for refusal.
Recruitment Lessons Learned
Overall, the CTEPP recruitment methods worked well.
However, recruitment required more time and effort than
anticipated. In CTEPP the problem was exacerbated by the
Census 2000 requirement that no field studies involving
contact with individuals could be conducted during the
census. As a result, subject recruitment had to be deferred for
about 4 months. Additionally, many study participants did
indicate that they should receive greater compensation for
conducting the burdensome data collection activities. In
future, similar studies, the recruitment success can be
enhanced by implementing several lessons learned from the
CTEPP study, as discussed below.
Obtaining consent from the day-care center director is the
key to the success of recruiting subjects from the day-care
sampling frame. If the day-care center is a national franchise,
the corporate headquarters or its regional director usually
makes the final decision. Their main concerns are staff and
time constraints, any potential liability caused by participa-
tion, and potential adverse impacts on day-care parents. It
was very time-consuming to contact the day-care directors.
For some day-care centers, it took several months for the
CTEPP staff to obtain a final decision from the day-care
directors.
Design and sampling methodogy for the CTEPP study Wilson et al.
Journal of Exposure Analysis and Environmental Epidemiology (2004) 14(3) 269
Once consent was obtained from the day-care director, the
next challenge was to recruit day-care parents. The study
staff had to obtain a list of day-care classrooms and age-
eligible children, select and identify children, and distribute
initial information packages to the parents of selected
children. Most centers would not allow study staff to contact
parents directly. Instead, parents were requested to call a
study toll-free number. Many did not respond.
Recommendations to improve day-care participation
include: (1) Increase the compensation to day-care centers,
both to the center director and to the individual classroom
teachers. (2) Prepare a special document, that would be
designed to ease the concerns of the day-care directors. This
document would include the following information, clearly
highlighted: guaranteed privacy, the compensation for time
spent on the project activities, a description of day-care
recruitment procedures and study activities, and the assis-
tance that would be provided by project staff. (3) Design and
implement a study website, which both explains the study
and also provides a means for potential participants to ask
questions. (4) Increase the staff and resources for the project
recruitment team, so that more intensive recruitment
activities, such as follow-up visits to the day-care centers,
can be conducted. (5) Increase the compensation to day-care
parents, to ensure that they are adequately compensated for
the real costs to them such as the cost of extra food for the
duplicate plate sampling. (6) Conduct additional in-depth
staff training on subject recruitment and data collection
activities. (7) For meetings with parents at the day-care
center, send at least two or three staff members. This would
ensure full attention by the staff to all participants and
minimize parents’ waiting time. (8) Minimize participant
burden as much as possible. Especially important here is
streamlining the child time–activity diaries, which most
caregivers and teachers found quite burdensome. Other
means of reducing burden were discussed earlier in this text,
under the recruitment section.
Recruitment through telephone screening, using the
county-by-county approach, worked very well. However,
the advance mailing of information about the study was not
cost-effective, as about 65% of the mailed packages were
returned due to bad addresses. Recommendations to improve
participation through telephone screening recruitment in-
clude: (1) Increase the compensation to the parents, to ensure
that they are adequately compensated for the real costs to
them, such as for extra food and utilities. (2) Mail study
brochures and introductory letter to the potential participant
immediately after completion of their telephone screening. (3)
Minimize participant burden as much as possible. Especially
important here is streamlining the child time–activity diaries,
which most caregivers found quite burdensome.
In summary, recruitment for the CTEPP study required
more effort and time than anticipated. As described earlier, it
was very time-consuming to obtain responses from the day-
care directors and parents. In future, similar studies, more
time and staff resources should be allocated to the
recruitment task. Recruitment should begin at least 6 months
before field sampling begins.
Problems Encountered in Field Sampling
A frequent problem encountered at the children’s day-care
centers was the teachers’ difficulty in recording the time–
activity diary for more than one child in a classroom.
Although field staff went over the time–activity diaries with
the adult participants in detail before sampling, the detail
required was overwhelming to some participants. As a result,
in some cases, the coverage of the time periods in the child
time–activity diaries was incomplete. In future studies, efforts
should be made to simplify such diaries, perhaps by
electronic recording or other improved methods. Teachers
were also reluctant to obtain and store children’s urine
samples for later pick-up. To remedy these difficulties, if
requested, field staff came to the centers and assisted as
needed in sample collection.
At first, parents sometimes had difficulty understanding
the requirements for duplicate diet collection and recording
the time–activity diaries, although they had previously
received individual training sessions with field staff before
the start of sampling. To remedy this problem, each aspect of
the study, including collection of duplicate food, urine, and
hand-wipe samples, and the interviews and activity diaries,
was explained in detail early in the recruitment period, and
again in the pre-sampling training visit.
Communication issues in the field were related to problems
with directions, equipment malfunctions, and scheduling
changes. Participants sometimes needed to contact the field
staff with questions or for assurance during the sampling
period. To alleviate these problems, all field staff were
furnishedwithcellphones,andcouldbecontactedatall
times by other staff or by participants themselves.
Completion Rates
Overall, more than 8200 discrete personal and environmental
samples, including field blanks, were collected in the two
states. In addition, pre- and postmonitoring questionnaires,
and child time–activity diaries and food surveys were
completed for each of the 257 participating children, at
home and at day care. Videotaping was completed for 100%
of those participants selected. Despite the high burden of the
study, only two participants dropped out during sampling F
one because of opposition from the landlord, and the other
because of opposition from the spouse. This 99 þ%
completeness compares favorably with study completeness
goals of 85%.
Distribution of Participants
The eventual distributions of the CTEPP participants among
low-income and middle/high-income, urban and rural, are
Design and sampling methodogy for the CTEPP studyWilson et al.
270 Journal of Exposure Analysis and Environmental Epidemiology (2004) 14(3)
shown in Figures 5 (NC) and 6 (OH). The project was
successful in oversampling low-income participants. In NC,
60% of day-care participants and 31% of nonday-care
participants were classified as low-income, overall about 45%
compared with the target of 38%, due to the larger number
of low-income families in the day-care stratum. In OH, 50%
of day-care participants and 26% of nonday-care partici-
pants were classified as low-income, overall about 38%, right
on target. By oversampling for low-income, using probability
of selection proportional to the size of the overall population
(PPS), as discussed earlier, one result was some oversampling
of urban participants. In NC, 86% of day-care participants
and 82% of nonday-care participants were classified as
urban, compared to 60% urban in the general NC
population. In OH, 86% of day-care participants and 87%
of nonday-care participants were classified as urban,
compared to 77% urban in the general OH population.
Survey weights will be calculated to account for oversampling
of certain subpopulations and for nonresponse bias when
calculating population-based estimates.
Conclusions
The CTEPP study of the aggregate exposures of preschool
children and their primary caregivers to persistent organic
pollutants in their everyday environments was successful in
recruiting and completing field sampling for children and
adults in 257 households in NC and OH. The sampling
design allowed probability-based sampling, while allowing
sufficient oversampling of particular groups for testing the
study hypotheses. The recruiting methods were successful at
enrolling and retaining subjects in the study, despite the high
burden put on participants. The field sampling methods were
successful at obtaining more than 8200 samples, which when
analyzed will provide valuable information on children’s
everyday exposures, and those of their adult primary
caregivers. The survey methods were successful at eliciting
extensive information on exposure-related behavior and on
sources of exposure in the children’s environments, which will
be exceptionally useful when incorporated into the study
database.
The implementation of such a large study requires lengthy
preparation, detailed protocols, meticulous attention to
detail, and dedicated, well-trained staff. It is necessary to
establish rapport with the participants and ensure that they
are treated as partners in the research.
Despite the high burden put on participants, many were
willing to cooperate in the research and provide extensive
information, knowing that their contributions to children’s
health and future are valuable. The success of the design and
field effort of the CTEPP study is evidenced by the
accomplishment of recruiting 91–108% of the planned
households, the high completion rates, the achievement of
study objectives relative to hypothesis testing, and the
willingness of 95% of the CTEPP participants to be
contacted in the future for other, similar studies.
Acknowledgments
Many persons contributed to the design and field activities of
CTEPP, including the individuals mentioned specifically
below. For their contributions to the recruitment and field
activities, we thank Suzanne Anderson, Carey Aselage,
Suzanne Benny, Chad Book, Tori Branch, Brigette Brevard,
Lisa Bryant, Martha Chapman, John Cashwell, Fred Crane,
Carla Dagnino, Lauren DiBiase, Josh Finegold, Stephanie
Low
60%
Middle/High
35%
Unknown
5%
NC Family Income
Day Care Sample
Urban
86%
Rural
14%
NC Urbanicity
Day Care Sample
Low
31%
Middle/High
63%
Unknown
6%
NC Family Income
Non-Day Care Sample
Urban
82%
Rural
18%
NC Urbanicity
Non-Day Care Sample
Figure 5. Family income and urbanicity characteristics of NC
participants.
Low
50%
Middle/High
41%
Unknown
9%
OH Family Income
Day Care Sample
Urban
86%
Rural
14%
OH Urbanicity
Day Care Sample
Low
26%
Middle/High
73%
Unknown
1%
OH Family Income
Non-Day Care Sample
Urban
87%
Rural
13%
OH Urbanicity
Non-Day Care Sample
Figure 6. Family income and urbanicity characteristics of OH
participants.
Design and sampling methodogy for the CTEPP study Wilson et al.
Journal of Exposure Analysis and Environmental Epidemiology (2004) 14(3) 271
Gray, LaTisha Griffin, Sherry Hubbard, Robyn Imm, Leslie
Lantz, Donna Magbag, Greg McDevitt, James McDonell,
Frances Patterson, Jan Satola, Eiko Weller, Leslie Wilson,
and Susan Winnard. For their insightful suggestions on the
design of the study, we thank Gary Evans, Robert Lewis,
Thomas McCurdy, Elaine Cohen Hubal, Maurice Berry,
Jim Quackenboss, Elizabeth Betz, and Carvin Stevens. For
their valuable comments on the study design, we thank
Maria Morandi, Ross Leidy, and Natalie Freeman.
Additionally, we thank Robert Lordo for his helpful review
and comments on survey sampling.
The United States Environmental Protection Agency
through its Office of Research and Development funded
and managed the research described here under Contract
#68–D99–011 to Battelle. It has been subjected to Agency
review and approved for publication. Mention of trade
names or commercial products does not constitute an
endorsement or recommendation for use.
References
Adgate J.L., Barr D.B., Clayton C.A., Eberly L.E., Freeman N.C.,
Lioy P.J., Needham L.L., Pellizzari E.D., Quackenboss J.J., Roy
A., and Sexton K. Measurement of children’s exposure to pesticides:
analysis of urinary metabolite levels in a probability-based sample.
Environ Health Perspect 2001: 109: 583–590.
Adgate J.L., Clayton C.A., Quackenboss J.J., Thomas K.W.,
Whitmore R.W., Pellizzari E.D., Lioy P.J., Shubat P., Stroebel
C., Freeman N.C.G., and Sexton K. Measurement of multi-
pollutant and multi-pathway exposures in a probability-based sample
of children: practical strategies for effective field studies. J. Expos
Anal Environ Epidemiol 2000: 10: 650–661.
Acheson K.J., Campbell I.T., Edholm O.G., Miller D.S., and Stock
M.J. The measurement of food and energy intake in man Fan
evaluation of some techniques. Am J Clin Nutr 1980: 33: 1147–1154.
Akland G.G., Pellizzari E.D., Hu Y., Roberds M., Rohrer C.A., Leckie
J.O., and Berry M.R. Factors influencing total dietary exposures of
young children. J Expos Anal Environ Epidemiol 2000: 10: 710–722.
ASTM FStandard practice for collection of floor dust for chemical
analysis D5438-94. Annual Book of ASTM Standards. Vol. 11.03.
American Society for Testing and Materials, West Conshohoken,
PA, 1997, pp. 517–523.
Baker J.A., Mendola P., Barr D., Walsh D., Creason J., and Needham
L. Non- residential organophosphorus pesticide use as a predictor of
children’s urinary metabolite levels. Presented at the Joint Meeting of
the International Society of Exposure Analysis and the International
Society for Environmental Epidemiology Vancouver, BC, Canada,
August, 2002.
Block G. A review of validations of dietary exposure assessment
methods. Am J Epidemiol 1982: 115: 492–505.
Brock J.W., Caudill S.P., Silva M.J., Needham L.L., and Hilborn E.D.
Phthalate monoesters levels in the urine of young children. Bull
EnvironContaminToxicol2002: 68: 309–314.
Buckley B., Ettinger A., Hore P., Lioy P., and Freeman N. Using
observational information in planning and implementation of field
studies with children as subjects. J Expos Anal Environ Epidemiol
2000: 10: 695–702.
Camann D.E., Harding H.J., Geno P.W., and Agrawal S.R.
Comparison of Methods to Determine Dislodgeable Residue Transfer
from Floors. EPA 600/R-96/089, 1996.
Chuang J.C., Callahan P.J., Lyu C.W., and Wilson N.K. Polycyclic
aromatic hydrocarbon exposures of children in low-income families.
J Expos Anal Environ Epidemiol 1999a: 9: 85–98.
Chuang J.C., Callahan P.J., Menton R.G., Gordon S.M., Lewis R.G.,
and Wilson N.K. Monitoring methods for polycyclic aromatic
hydrocarbons and their distribution in house dust and track-in soil.
Environ Sci Technol 1995: 29: 494–500.
Chuang J.C., Gordon S.M., Roberts J.W., Han W., and Ruby M.G.
Evaluation of HVS3 sampler for sampling polycyclic aromatic
hydrocarbons and polychlorinated biphenyls. EPA 600/R-94/188,
1994.
Chuang J.C., Hannan S.W., and Wilson N.K. Field comparison of
polyurethane foam and XAD2 resin for air sampling for
polynuclear aromatic hydrocarbons. Environ Sci Technol 1987: 21:
798–804.
Chuang J.C., Lyu C., Chou Y.L., Callahan P.J., Nishioka M.,
Andrews K., Pollard M.A., Brackney L., Hines C., Davis D.B.,
and Menton R. Evaluation and Application of Methods
for Estimating Children’s Exposure to Persistent Organic
Pollutants in Multiple Media. EPA 600/R-98/164a, 164b, and
164c. 1998.
Chuang J.C., Mack G.A., Kuhlman M.R., and Wilson N.K. Polycyclic
aromatic hydrocarbons and their derivatives in indoor and outdoor
air in an eight-home pilot study. Atmos Environ 1991: 25B:
369–380.
Chuang J.C., Wilson N.K., and Lewis R.G. Methodology of ambient
monitoring for polycyclic aromatic hydrocarbons. Fresenius Environ
Bull 1999b: 8: 547–556.
Cohen Hubal E.A., Sheldon L.S., McCurdy T.R., Rigas M.L., Burke
J.M., and Zartarian V.G. Children’s exposure assessment: a review
of factors influencing children’s exposure, and the data available to
characterize and assess that exposure. Environ Health Perspect
2000a: 108(6): 475–486.
Cohen Hubal E.A., Sheldon L.S., Zufall M.J., Burke J.M., and
Thomas K.W. The challenge of assessing children’s residential
exposure to pesticides. J Expos Anal Environ Epidemiol 2000b: 10:
638–649.
Duggan M.J., Inskip M.J., Rudle S.A., and Moorcroft J.S. Lead in
playground dust and on the hands of schoolchildren. Sci Total
Environ 1985: 44: 65–79.
Federal Register, Vol. 65, No. 33, February 15, 2000, pp. 7555–7557.
http://aspe.hhs.gov/poverty/.
Fennema O.R., and Anderson S.A., In: Macdonald I., (Ed.) Monitoring
Dietary Intakes. Springer, Berlin, 1991, p. 45.
Fenske R.A., Kedan G., Lu C., Fisker-Andersen J.A., and Curl C.L.
Assessment of organophosphate pesticide exposures in the diets of
preschool children in Washington State. J Expos Anal Environ
Epidemiol 2002a: 12: 21–28.
Fenske R.A., Kissel J.C., Lu C., Kalman D.A., Simcox N.J., Allen
E.H., and Keifer M.C. Biologically based pesticide dose estimates for
children in an agricultural community. Environ Health Perspect
2000a: 108: 515–520.
Fenske R.A., Lu C., Barr D., and Needham L. Children’s exposure
to chlorpyrifos and parathion in an agricultural community in
central Washington State. Environ Health Perspect 2002b: 110:
549–553.
Fenske R.A., Lu C., Simcox N.J., Loewenherz C., Touchstone J.,
Moate T.F., Allen E.H., and Kissel J.C. Strategies for assessing
children’s organophosphorus pesticide exposures in agricultural
communities. J Expos Anal Environ Epidemiol 2000b: 10:
662–671.
Fortune C.R. Round-robin Testing of Methods for Collecting Dislodge-
able Residues from Carpets. EPA 600/R-97/119, 1997.
FQPA. Food Quality Protection Act of 1996. http://www.epa.gov/
opppsps1/fqpa/fqpa-iss.htm.
Design and sampling methodogy for the CTEPP studyWilson et al.
272 Journal of Exposure Analysis and Environmental Epidemiology (2004) 14(3)
Freeman N.C., Jimenez M., Reed K.J., Gurunathan S., Edwards R.D.,
Roy A., Adgate J.L., Pellizzarri E.D., Quackenboss J., Sexton K.,
and Lioy P.J. Quantitative analysis of children’s microactivity
patterns: The Minnesota Children’s Pesticide Exposure Study.
J Expos Anal Environ Epidemiol 2001: 11: 501–509.
Freedman D.M., Stewart P., Kleinerman R.A., Wacholder S., Hatch
E.E., Tarone R.E., Robison L.L., and Linet M.S. Household solvent
exposures and childhood acute lymphoblastic leukemia. Am J Public
Health 2001: 91: 564–567.
Gurunathan S., Robson M., Freeman N., Buckley B., Roy A., Meyer
R., Bukowski J., and Lioy P.J. Accumulation of chlorpyrifos on
residential surfaces and toys accessible to children. Environ Health
Perspect 1998: 106: 9–16.
Heil M., Schilter B., Huggett A.C., and Haschke F. Toxicological
aspects of food for infants and children. Monatsschr Kinderheilk
1996: 144 (Suppl. 2): S224–S229.
Heudorf U., and Angerer J. Internal exposure to PAHs of children and
adults living in homes with parquet flooring containing high levels of
PAHs in the parquet glue. Int J Occup Environ Health 2001: 74:
91–101.
Karmaus W., DeKoning E.P., Kruse H., Witten J., and Osius N. Early
childhood determinants of organochlorine concentrations in school-
aged children. Pediatr Res 2001: 50: 322–323.
Koch D., Lu C., Fisker-Andersen J., Jolley L., and Fenske R.A.
Temporal association of children’s pesticide exposure and agricultural
spraying: Report of a longitudinal biological monitoring study.
Environ Health Perspect 2002: 110: 829–833.
Landrigan P.J. Pesticides and polychlorinated biphenyls (PCBs): an
analysis of the evidence that they impair children’s neurobehavioral
development. Mol Genet Metab 2001: 73: 11–17.
Landrigan P.J., Claudio L., Markowitz S.B., Berkowitz G.S., Brenner
B.L., Romero H., Wetmer J.G., Matte T.D., Gore A.C., Godbold
J.H., and Wolff M.S. Pesticides and inner-city children: exposures,
risks, and prevention. Environ Health Perspect 1999: 107(Suppl. 3):
431–437.
Lewis R.G., Fortmann R.C., and Camann D.E. Evaluation of methods
for monitoring the potential exposure of small children to pesticides
in the residential environment. Arch Environ Contamin Toxicol 1994:
26: 37–46.
Lu C., Knutson D.E., Fisker-Andersen J., and Fenske R.A. Biological
monitoring survey of organophosphorus pesticide exposures among
pre-school children in the Seattle metropolitan area. Environ Health
Perspect 2001: 109: 299–303.
Lyu C., Wilson N.K., Chuang J.C., Satola J., and Morgan M.K.
Recruitment and field sampling in the CTEPP North Carolina and
Ohio field studies. Presented in the Symposium on Young Children’s
Exposures to Persistent Organic Pollutants at the joint meeting of the
International Society of Exposure Analysis and the International
Society for Environmental Epidemiology, Vancouver, BC, Canada,
August, 2002.
Mills P.K., and Zahm S.H. Organophosphate pesticide residues in urine
of farmworkers and their children in Fresno County, California.
Am J Ind Med 2001: 40: 571–577.
Mukerjee D. Assessment of risk from multimedia exposures of children
to environmental chemicals. J Air Waste Manag Assoc 1998: 48:
483–501.
Needham L.L., and Sexton K. Assessing children’s exposure to
hazardous environmental chemicals: an overview of selected research
challenges and complexities. J Expos Anal Environ Epidemiol 2000:
10: 611–629.
Nishioka M.G., Lewis R.G., Brinkman M.C., Burkholder H.M., Hines
C.E., and Menkedick J.R. Distribution of 2,4-D in air and on
surfaces inside residences after lawn applications: comparing
exposure estimates from various media for young children. Environ
Health Perspect 2001: 109: 1185–1191.
O’Rourke M.K., Lizardi P.S., Rogan S.P., Freeman N.C., Aguirre A.,
and Saint C.G. Pesticide exposure and creatinine variation
among young children. J Expos Anal Environ Epidemiol 2000: 10:
672–681.
Perera F.P., Illman S.M., Kinney P.L., Whyatt R.M., Kelvin E.A.,
Shepard P., Evans D., Fullilove M., Ford J., Miller R.L.,
Meyer I.H., and Rauh V.A. The challenge of preventing envir-
onmentally related disease in young children: community-
based research in New York City. Environ Health Perspect 2002:
110: 197–204.
Raymer J.H., Akland G., Hu Y., Spruill M., Briggs K., and Childs B.
Study of exposure and body burden of children of different ages to
pesticides in the environment. Presented in the Symposium on Young
Children’s Exposures to Persistent Organic Pollutants at the Joint
Meeting of the International Society of Exposure Analysis and the
International Society for Environmental Epidemiology, Vancouver,
BC, Canada, August, 2002.
Rigas M.L., Okino M.S., and Quackenboss J.J. Use of a pharmaco-
kinetic model to assess chlorpyrifos exposure and dose in children,
based on urinary biomarker measurements. To x i col S c i 2001: 61:
374–381.
Stanek III E.J., and Calabrese E.J. Daily estimates of soil ingestion in
children. Environ Health Perspect 1995: 103: 276–285.
Tulve N.S., Suggs J.C., McCurdy T., Cohen Hubal E.A., and Moya J.
Frequency of mouthing behavior in young children. J Expos Anal
Environ Epidemiol 2002a: 12: 259–264.
Tulve N.S., Whitaker D., Fortmann R., Brown H., Bozeman E.R.,
Hilliard A., and Naeher L.P. Environmental measurements
of organophosphate and pyrethroid pesticides to assess exposures
of young children living in Jacksonville, FL. Presented in
the Symposium on Young Children’s Exposures to Persistent
Organic Pollutants at the Joint Meeting of the International
Society of Exposure Analysis and the International Society for
Environmental Epidemiology, Vancouver, BC, Canada, August,
2002b.
US Census Bureau. Current Population Survey 1960–2002. Annual
Demographic Supplements, and Census 2000 Summary File, 2002.
http://factfinder.census.gov/.
Vonmanikowsky S., Sagunski H., Papke O., Fabig K.R., Neus H.,
Schumann F., Spannhake I., Karmaus W., Osius N., and Zier B.
PCDD/PCDF indoor exposure in day-care centers and PCDD/
PCDF blood concentrations of female employees. Environ Health
Perspect 1998: 106(Suppl. 2): 707–714.
WIC (Women, Infants, and Children). 2000. http://www.fns.usda.gov/
wic/.
Wilhelm M., Schrey P., Wittsiepe J., and Heinzow B. Dietary intake of
persistent organic pollutants (POPs) by German children using
duplicate portion sampling. Int J Hyg Environ Health 2002: 204:
359–362.
Wilson N.K., Chuang J.C., and Kuhlman M.R. Sampling polycyclic
aromatic hydrocarbons and related semivolatile organic compounds
in indoor air. Indoor Air 1991: 4: 513–521.
Wilson N.K., Chuang J.C., and Lyu C. Evaluation of field
methods for estimating exposure of children in low-income
families to polycyclic aromatic hydrocarbons. Measurement
of Toxic and Related Air Pollutants. VIP-64: Air and Waste
Management Association, Pittsburgh, PA,, 1996: (VIP-64):
797–802.
Wilson N.K., Chuang J.C., and Lyu C. Multimedia concentrations of
PAH in several day care centers. Polycyclic Aromatic Compds 1999:
17: 255–265.
Design and sampling methodogy for the CTEPP study Wilson et al.
Journal of Exposure Analysis and Environmental Epidemiology (2004) 14(3) 273
Wilson N.K., Chuang J.C., and Lyu C. PAH exposures of nine
preschool children. Polycyclic Aromatic Compds 2000: 21:
247–259.
Wilson N.K., Chuang J.C., and Lyu C. Levels of persistent organic
pollutants in several child day care centers. J Expos Anal Environ
Epidemiol 2001: 11: 449–458.
Wilson N.K., Chuang J.C., Lyu C., Menton R., and Morgan M.K.
Aggregate exposures of nine preschool children to persistent organic
pollutantsatdaycareandathome.J Expos Anal Environ Epidemiol
2003: 13: 187–202.
Design and sampling methodogy for the CTEPP studyWilson et al.
274 Journal of Exposure Analysis and Environmental Epidemiology (2004) 14(3)
